CML (Continuous Machine Learning)

Open-source CI/CD for Machine Learning, automating ML workflows with Git.

by Iterative (DVC) · Marketing Automation

Executive Summary

CML (Continuous Machine Learning) is an open-source tool designed for implementing continuous integration and continuous delivery (CI/CD) practices in machine learning projects. It enables teams to automate their ML workflows by integrating seamlessly with Git, allowing for version control of code, data, and models. CML leverages cloud infrastructure for remote experiments and provides automatic reporting capabilities, including graphs and images. It works in conjunction with DVC (Data Version Control), which offers robust data and model versioning, experiment tracking, and a powerful Python API for accessing data and logging metrics and artifacts. The DVC 3.0 stack extends beyond command-line operations, offering a unified view of all files within a repository—whether Git-tracked, DVC-tracked, or untracked. This allows for smarter interaction with various cloud and remote storage solutions, streamlining the entire ML lifecycle from notebook exploration to comprehensive model management.

Use Cases

  • Implementing CI/CD for machine learning projects
  • Automating ML workflows with Git
  • Leveraging cloud infrastructure for remote ML experiments
  • Automatic reporting of ML metrics, graphs, and images
  • Data and model versioning and management
  • Experiment tracking and logging metrics, plots, and models
  • Automating data validation and model monitoring pipelines

Features

Visibility

  • Experiment Tracking & Visualization: Log and visualize metrics, plots, and models from ML experiments to compare runs and track performance.
  • Automatic Reporting: Generate reports including graphs and images automatically within CI/CD pipelines.

Technical Specifications

Architecture
Open-source, Git-based architecture for versioning data, models, and code. Integrates with various cloud/remote storage solutions and provides a Python API for programmatic interaction.
Deployment
Self-Hosted
API Available
Yes

Infrastructure

  • AWS
  • GCP
  • Azure

Integrations

  • Git
  • AWS S3
  • GCP GCS
  • Azure Blob Storage

Security & Compliance

Encryption: Leverages encryption capabilities of underlying cloud/remote storage providers.

Pricing

Model
Free (open-source)
Starting Price
Free
Target Customer
SMB,Mid-Market,Enterprise
Free Trial
Yes, Unlimited (no credit card required)